Communication Apparatus, Data Set Providing Apparatus, Method for Training AI/ML Model, and Method for Providing Information on Which to Base Learning of AI/ML Model

The present disclosure relates to a communication apparatus, a data set providing apparatus, a method for training an AI/ML model, and a method for providing information on which to base learning of an AI/ML model.

Today, Beyond 5G and 6G are under discussion as next-generation mobile communication systems in the 3rd Generation Partnership Project.

Further improvements in enhanced mobile broadband (eMBB), massive machine type communications (mMTC), and ultra-reliable and low-latency communications (URLLC) are expected in the Beyond 5G and 6G radio access schemes. In order to achieve such improvements, the use of an artificial intelligence/machine learning (AI/ML) model to process a signal transmitted and received through wireless communication is under discussion.

Non-Patent Literature 1: R1-2203280, “General aspects of AI PHY”, Ericsson, 3GPP TSG-RAN WG1 Meeting #109-e, Electronic Meeting, May 16-27, 2021

When, for example, signal processing is performed between an AI/ML model of a terminal apparatus and an AI/ML model of a base station, there is a possibility that the signal processing between the respective AI/ML models will not be performed properly if specifications or implementations of the respective AI/ML models are different from each other.

Further, when, for example, signal processing is performed between an AI/ML model of each one of a plurality of terminal apparatuses and an AI/ML model of a base station, there is a possibility that the signal processing between the terminal apparatus and the base station will not be performed properly if specifications or implementations of the respective AI/ML models of the plurality of terminal apparatuses are different from each other due to, for example, a difference in terminal vendor.

The present disclosure addresses the issues described above, and it is an object of the present disclosure to provide a communication apparatus, a data set providing apparatus, a method for training an AI/ML model, and a method for providing information on which to base learning of an AI/ML model that make it possible to perform signal processing without being aware of a difference in AI/ML model.

A communication apparatus according to the present disclosure includes a signal processor that includes an AI/ML model; a receiver that receives, from another communication apparatus, information on which to base learning; and a controller that sets the AI/ML model on the basis of the information on which to base learning.

Further, a data set providing apparatus according to the present disclosure is a data set providing apparatus that is accessible in common by communication apparatuses of different vendors, the data set providing apparatus including a storage that stores therein information on which to base learning of an AI/ML model; a receiver that receives, from a communication apparatus, a request for provision of the information on which to base learning of the AI/ML model; and a transmitter that transmits, to the communication apparatus, the information on which to base learning of the AI/ML model.

Furthermore, a method for training an AI/ML model according to the present disclosure includes receiving information on which to base learning; and setting the AI/ML model on the basis of the information on which to base learning.

Further, a method for providing information on which to base learning of an AI/ML model according to the present disclosure includes receiving a request for provision of the information on which to base learning; and transmitting the information on which to base learning to a transmission source that has transmitted the request for the provision.

Embodiments of the present disclosure will now be described in detail below with reference to the drawings. In the figures, similar or corresponding structural elements will be denoted by the same reference symbol, and detailed descriptions thereof are omitted as appropriate.

1 FIG. 1 FIG. 1 1 10 20 30 40 1 1 20 30 40 illustrates a configuration of a wireless communication systemaccording to a first embodiment of the present disclosure. The wireless communication systemincludes a management apparatus, a base station, a relay station, and a terminal apparatus. The wireless communication systemprovides a user with a wireless network that enables mobile communication by wireless communication apparatuses included in the wireless communication systemoperating cooperatively. The wireless network of the first embodiment includes a radio access network RAN and a core network CN. In the first embodiment, the wireless communication apparatus is an apparatus that includes a function of wireless communication, and the base station, the relay station, and the terminal apparatuseach correspond to the wireless communication apparatus in the example illustrated in.

1 10 20 30 40 1 10 10 10 20 20 20 20 1 30 30 30 40 40 40 40 1 FIG. a b a b c a b a b c The wireless communication systemmay include a plurality of management apparatuses, a plurality of base stations, a plurality of relay stations, and a plurality of terminal apparatuses. In the example illustrated in, the wireless communication systemincludes management apparatusesandthat correspond to the management apparatuses, and includes base stations,, andthat correspond to the base stations. Further, the wireless communication systemincludes relay stationsandthat are the relay stations, and includes terminal apparatuses,, andthat are the terminal apparatuses.

1 FIG. Each of the wireless communication apparatuses illustrated inmay be considered an apparatus in a logical sense. In other words, some of the respective wireless communication apparatuses may be implemented by, for example, virtual machines (VMs), containers, or dockers and respectively mounted on physically identical pieces of hardware.

1 40 20 The wireless communication systemmay support a radio access technology (RAT) such as Long Term Evolution (LTE) or New Radio (NR). Each of the LTE and the NR is a type of cellular communication technology, and enables the terminal apparatusto perform mobile communication by arranging, in the form of cells, a plurality of areas covered by the base station.

1 A radio access scheme of the wireless communication systemis not limited to, for example, the LTE or the NR, and may be another radio access scheme such as wideband code-division multiple access (W-CDMA) or code-division multiple access 2000 (cdma2000).

20 30 1 1 The base stationand relay stationincluded in the wireless communication systemmay be a ground station or a non-ground station. The non-ground station may be a satellite station or an aircraft station. When the non-ground station is a satellite station, the wireless communication systemmay be a bent-pipe (transparent) mobile satellite communication system.

In the first embodiment, the ground station (also referred to as ground base station) refers to a base station (including a “relay station”) that is installed on the ground. Here, the term “on the ground” has a broad meaning, and includes not only “on land” but also “under the ground”, “on water”, and “in water”. Note that, in the following description, the term “ground station” may be replaced with the term “gateway”.

Note that a base station of the LTE may be referred to as an evolved Node B (eNodeB) or an eNB. Further, a base station of the NR may be referred to as a gNodeB or a gNB. In the LTE and the NR, a terminal apparatus (also referred to as a “mobile station” or a “terminal”) may be referred to as user equipment (UE).

40 20 30 In the first embodiment, the concept of the wireless communication apparatus includes not only a portable mobile object apparatus (a terminal apparatus) such as a mobile terminal but also an apparatus installed in a structure or a mobile object. The structure or mobile object itself may be considered the wireless communication apparatus. Further, the concept of the wireless communication apparatus includes not only the terminal apparatusbut also the base stationand the relay station. The wireless communication apparatus is a type of processing apparatus or a type of information processing apparatus. The wireless communication apparatus can also alternatively be phrased as a transmission apparatus or a reception apparatus.

1 Configurations of wireless communication apparatuses that are included in the wireless communication systemare specifically described below. Note that the configurations of the respective wireless communication apparatuses described below are merely examples. The configurations of the wireless communication apparatuses may be different from configurations describe below.

10 10 20 10 10 10 10 10 The management apparatusis an apparatus that manages a wireless network. For example, the management apparatusis an apparatus that manages communication performed by the base station. The management apparatusincludes a function of, for example, a mobility management entity (MME) when the core network CN is an evolved packet core (EPC). The management apparatusincludes a function of, for example, an access and mobility management function (AMF) and/or of a session management function (SMF) when the core network CN is a 5G core network (5GC). Note that the function of the management apparatusis not limited to the MME, the AMF, and the SMF. The management apparatusmay be an apparatus that includes a function of a network slice selection function (NSSF), an authentication server function (AUSF), or unified data management (UDM). The management apparatusmay be an apparatus that includes a function of a home subscriber server (HSS).

10 10 10 10 10 The management apparatusmay include a function of a gateway. The management apparatusmay include a function of a serving gateway (S-GW) or a packet data network gateway (P-GW) when the core network CN is an EPC. The management apparatusmay include a function of a user plane function (UPF) when the core network CN is a 5GC. The management apparatusdoes not necessarily have to be an apparatus that is included in the core network CN. The management apparatusmay be an apparatus that serves as a radio network controller (RNC) when the core network CN is a core network of wideband code-division multiple access (W-CDMA) or code-division multiple access 2000 (cdma2000).

2 FIG. 2 FIG. 2 FIG. 10 10 11 12 13 10 10 illustrates a configuration of the management apparatusaccording to the first embodiment. The management apparatusincludes a communication section, a storage, and a controller. Note thatillustrates a functional configuration, and a hardware configuration may be different from the configuration illustrated in. Further, functions of the management apparatusmay be statically or dynamically distributed to be provided to a plurality of physically separated configurations. The management apparatusmay include a plurality of server apparatuses.

11 20 30 11 11 11 11 10 11 13 The communication sectionis a communication interface used to communicate with a wireless communication apparatus (such as the base stationor the relay station). The communication sectionmay be a network interface or an apparatus connection interface. The communication sectionmay be a local area network (LAN) interface such as a network interface card (NIC), or may be a Universal Serial Bus (USB) interface including, for example, a USB host controller or a USB port. The communication sectionmay be a wired interface or a wireless interface. The communication sectionserves as a communication device of the management apparatus. The communication sectionis controlled by the controller.

12 12 10 12 40 12 40 12 40 The storageis a data readable/writable storage apparatus such as a dynamic random access memory (DRAM), a static random access memory (SRAM), a flash memory, or a hard disk. The storageserves as a storage device of the management apparatus. For example, the storagestores therein a connection state of the terminal apparatus. The storagestores therein a state of radio resource control (RRC) of the terminal apparatusand a state of EPS connection management (ECM) or 5G system connection management (CM). The storagemay serve as a home memory that stores therein position information regarding a position of the terminal apparatus.

13 10 13 13 10 13 The controlleris a controller that controls each component of the management apparatus. The controllermay be implemented by a processor such as a central processing unit (CPU) or a micro processing unit (MPU). In particular, the controllermay be implemented by a processor executing, using a random access memory (RAM) or the like as a working region, various programs stored in a storage apparatus situated inside of the management apparatus. The controllermay be implemented by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). Any of the CPU, the MPU, the ASIC, and the FPGA can be considered a controller.

20 40 20 40 30 40 The base stationis a wireless communication apparatus that wirelessly communicates with the terminal apparatus. The base stationmay wirelessly communicate with the terminal apparatusthrough the relay station, or may wirelessly communicate with the terminal apparatusdirectly.

20 20 20 20 20 The base stationis an apparatus that corresponds to a wireless base station (such as a base station, Node B, eNB, or gNB) or a wireless access point (an access point). The base stationmay be a wireless relay station. The base stationmay be an optical extension apparatus called a remote radio head (RRH). The base stationmay be a reception station such as a field pickup unit (FPU). The base stationmay be an integrated access and backhaul (IAB) donor node or IAB relay node that provides a wireless access link and a wireless backhaul link using time division multiplexing, frequency division multiplexing, or spatial division multiplexing.

20 20 20 20 20 20 20 A wireless access technology used by the base stationmay be a cellular communication technology. The wireless access technology used by the base stationmay be a wireless LAN technology. The wireless access technology used by the base stationmay be a low-power wide-area (LPWA) communication technology. Note that the wireless access technology used by the base stationis not limited thereto, and may be another wireless access technology. Wireless communication used by the base stationmay be wireless communication using millimeter waves. The wireless communication used by the base stationmay be wireless communication using radio waves, or wireless communication using infrared light or visible light, that is, optical wireless communication. The base stationmay be capable of

40 20 20 performing non-orthogonal multiple access (NOMA) communication with the terminal apparatus. The NOMA communication refers to communication (transmission, reception, or both of them) using non-orthogonal resources. The base stationmay be capable of performing the NOMA communication with another base station.

20 20 The base stationsmay be capable of communicating with the core network CN through an interface between the base station and the core network CN, such as an S1 interface. This interface may be a wired or wireless interface. The base stationmay be capable of communicating with another base station through an interface between base stations, such as an X2 interface. This interface may be a wired or wireless interface.

The concept of the base station (also referred to as a “base station apparatus”) includes not only a donor base station but also a relay base station (also referred to as a “relay station”). The concept of the base station includes not only a structure including a function of the base station but also an apparatus installed in the structure.

Examples of the structure include buildings such as a high-rise building, a house, a steel tower, station facilities, airport facilities, harbor facilities, and a stadium. The concept of the structure includes not only buildings but also non-building structures such as a tunnel, a bridge, a dam, a wall, and an iron pillar; and facilities such as a crane, a gate, and a windmill. The concept of the structure includes not only structures on land (on the ground in a narrow sense) or under the ground but also structures on water such as a pier or a megafloat, and structures in water such as marine observation facilities. The base station can also alternatively be phrased as an information processing apparatus.

20 20 20 20 The base stationmay be a fixed station, or a movable wireless communication apparatus, that is, a mobile station. The base stationmay be an apparatus installed in a mobile object, or may be the mobile object itself. A relay station that has the mobility can be considered the base stationserving as a mobile station. An apparatus, such as a vehicle, an unmanned aerial vehicle (UAV) represented by a drone, or a smartphone, that originally has the mobility and includes at least a portion of a function of a base station, can also be considered the base stationserving as a mobile station.

The mobile object may be a mobile terminal such as a smartphone or a mobile phone. The mobile object may be a mobile object (a vehicle such as an automobile, a bicycle, a bus, a truck, a motorcycle, a train, and a linear motor car) that moves on land (on the ground in a narrow sense), or a mobile object (such as a subway) that moves under the ground (for example, in a tunnel).

The mobile object may be a mobile object (such as vessels including a passenger ship, a cargo ship, and a hovercraft) that moves on water, or a mobile object (such as submersible ships including a submersible, a submarine, and an unmanned submersible machine) that moves in water.

The mobile object may be a mobile object (such as aircraft including an airplane, an airship, and a drone) that moves in the atmosphere of the earth.

20 20 20 20 20 1 20 The base stationmay be a ground base station (a ground station) installed on the ground. The base stationmay be a base station provided to a structure on the ground, or may be a base station installed in a mobile object that moves on the ground. The base stationmay be an antenna installed in a structure such as a building, and a signal processing apparatus connected to the antenna. The base stationmay be the structure or mobile object itself. The term “on the ground” has a broad meaning, and includes not only “on land” (“on the ground” in a narrow sense) but also “under the ground”, “on water”, and “in water”. The base stationis not limited to a ground base station. When the wireless communication systemis a satellite communication system, the base stationmay be an aircraft station. As viewed from a satellite station, an aircraft station situated on the earth is a ground station.

20 20 20 The base stationis not limited to a ground station. The base stationmay be a non-ground base station (a non-ground station) that can float in the air or in space. The base stationmay be an aircraft station or a satellite station. The satellite station is a satellite station that can float outside of the atmosphere of the earth. The satellite station may be an apparatus that is included in a space mobile object such as an artificial satellite, or may be the space mobile object itself. The space mobile object is a mobile object that moves outside of the atmosphere of the earth. Examples of the space mobile object include artificial celestial objects such as an artificial satellite, a spacecraft, a space station, and a probe.

A satellite serving as a satellite station may be any of a low earth orbiting (LEO) satellite, a medium earth orbiting (MEO) satellite, a geostationary earth orbiting (GEO) satellite, and a highly elliptical orbiting (HEO) satellite. The satellite station may be an apparatus that is included in a low earth orbiting satellite, a medium earth orbiting satellite, a geostationary earth orbiting satellite, or a highly elliptical orbiting satellite.

The aircraft station is a wireless communication apparatus, such as an aircraft, that can float in the atmosphere of the earth. The aircraft station may be an apparatus that is included in, for example, an aircraft, or may be the aircraft itself. The concept of the aircraft includes not only heavy aircraft such as an airplane and a glider but also light aircraft such as a balloon and an airship. The concept of the aircraft includes not only heavy aircraft and light aircraft but also rotorcraft such as a helicopter and an autogyro. The aircraft station, or an aircraft that includes the aircraft station may be an unmanned aircraft such as a drone.

The concept of the unmanned aircraft also includes an unmanned aircraft system (UAS) and a tethered UAS. The concept of the unmanned aircraft includes a lighter-than-air (LTA) UAS and a heavier-than-air (HTA) UAS. The concept of the unmanned aircraft also includes a high-altitude UAS platform (HAP).

20 20 20 20 The coverage of the base stationmay be as relatively large as a macrocell, or may be as relatively small as a picocell. The coverage of the base stationmay be as extremely small as a femtocell. The base stationmay include a beamforming function. For the base station, a cell or a service area may be formed for each beam.

3 FIG. 3 FIG. 3 FIG. 20 20 21 22 23 25 20 illustrates a configuration of the base stationaccording to the first embodiment. The base stationincludes a wireless communication section, a storage, a decoder, and a controller. Note thatillustrates a functional configuration, and a hardware configuration may be different from the configuration illustrated in. Further, functions of the base stationmay be distributed to be provided to a plurality of physically separated configurations.

21 30 40 20 21 25 21 21 21 21 The wireless communication sectionis a signal processor used to wirelessly communicate with another wireless communication apparatus (such as the relay station, the terminal apparatus, or another base station). The wireless communication sectionis controlled by the controller. The wireless communication sectionsupports at least one radio access scheme. The wireless communication sectionmay support both the NR and the LTE. The wireless communication sectionmay support, for example, W-CDMA and cdma2000 in addition to the NR and the LTE. The wireless communication sectionmay support an automatic retransmission technology such as a hybrid automatic repeat request (HARQ).

21 211 212 213 21 211 212 213 21 21 211 212 213 21 21 The wireless communication sectionincludes a transmitter, a receiver, and an antenna. The wireless communication sectionmay include a plurality of transmitters, a plurality of receivers, and a plurality of antennas. When the wireless communication sectionsupports a plurality of radio access schemes, the respective components of the wireless communication sectionmay be individually configured for each radio access scheme. The transmitterand the receivermay be individually configured for each of the LTE and the NR. The antennamay include a plurality of antenna elements such as a plurality of patch antennas. The wireless communication sectionmay include a beamforming function. The wireless communication sectionmay include a function of polarization beamforming that uses vertically polarized waves (V-polarized waves) and horizontally polarized waves (H-polarized waves).

211 211 24 The transmitterperforms processing of transmitting downlink control information and downlink data. For example, first, the transmitterencodes downlink control information and downlink data that are input from the controller, using an encoding scheme such as block encoding, convolutional encoding, or turbo encoding. Encoding using a polar code or encoding using a low density parity check (LDPC) code may be performed as the encoding.

211 Next, the transmittermodulates an encoded bit according to a specified modulation scheme such as BPSK, QPSK, 16QAM, 64QAM, or 256QAM. In this case, signal points on a constellation do not necessarily have to be equally spaced. In other words, the constellation may be a non-uniform constellation (NUC).

211 211 211 211 213 Next, the transmittermultiplexes a modulation symbol of each channel and a downlink reference signal, and provides the multiplexed modulation symbol and downlink reference signal to a specified resource element. Next, the transmitterperforms various signal processes on the multiplexed signal. For example, the transmitterperforms processes such as transformation into a frequency domain using fast Fourier transform, addition of a guard interval (cyclic prefix), generation of a baseband digital signal, conversion into an analog signal, quadrature modulation, up-conversion, removal of an extra frequency component, and power amplification. Finally, a signal generated by the transmitteris transmitted through the antenna.

212 213 212 The receiverprocesses an uplink signal received through the antenna. For example, first, the receiverperforms, on the uplink signal, down-conversion, removal of an unnecessary frequency component, control of an amplification level, quadrature demodulation, conversion into a digital signal, removal of a guard interval (cyclic prefix), extraction of a frequency domain signal using fast Fourier transform, and the like.

212 212 Next, from the signal on which these processes have been performed, the receiverseparates an uplink channel such as a physical uplink shared channel (PUSCH) and a physical uplink control channel (PUCCH), and an uplink reference signal. Next, the receiverdemodulates a received signal using a modulation symbol of the uplink channel according to a modulation scheme such as binary phase shift keying (BPSK) or quadrature phase shift keying (QPSK). The modulation scheme may be 16 quadrature amplitude modulation (QAM), 64QAM, or 256QAM. In this case, signal points on a constellation do not necessarily have to be equally spaced. In other words, the constellation may be a non-uniform constellation.

212 24 Next, the receiverperforms decoding processing on a demodulated encoded bit of the uplink channel. Finally, decoded uplink data and uplink control information are output to the controller.

213 213 213 213 21 21 213 213 21 21 The antennais an antenna apparatus that performs conversion between current and radio waves. The antennamay include a single antenna element, that is, for example, a single patch antenna. The antennamay include a plurality of antenna elements, that is, for example, a plurality of patch antennas. When the antennaincludes a plurality of antenna elements, the wireless communication sectionmay include a beamforming function. The wireless communication sectionmay be configured to generate a directional beam by controlling the directivity of a radio signal using a plurality of antenna elements. The antennamay be a dual-polarized antenna. When the antennais a dual-polarized antenna, the wireless communication sectionmay use vertically polarized waves (V-polarized waves) and horizontally polarized waves (H-polarized waves) upon transmitting a radio signal. The wireless communication sectionmay control the directivity of the radio signal transmitted using the vertically polarized waves and the horizontally polarized waves.

22 22 20 The storageis a data readable/writable storage apparatus such as a DRAM, an SRAM, a flash memory, or a hard disk. The storageserves as a storage device of the base station.

23 The decoderincludes a function of performing signal processing using an AI/ML model. The AI/ML model is a neural network model obtained by machine learning or deep learning. Examples of the neural network model may include a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, and a long short-term memory (LSTM) model. The AL/ML model may be any of these models, or may be obtained by combining these models in series or in parallel.

23 23 23 23 6 FIG. The decoderrestores and outputs first data by performing signal processing using an AI/ML model, where second data is used as input. The decodermay be implemented by a processor such as a CPU or an MPU. The decodermay be implemented by an integrated circuit such as an ASIC or an FPGA. The configuration of the decoderwill be described in further detail later with reference to.

25 20 25 25 20 25 25 The controlleris a controller that controls each component of the base station. The controllermay be implemented by a processor such as a CPU or an MPU. In particular, the controllermay be implemented by a processor executing, using a RAM or the like as a working region, various programs stored in a storage apparatus situated inside of the base station. The controllermay be implemented by an integrated circuit such as an ASIC or an FPGA. Any of the CPU, the MPU, the ASIC, and the FPGA can be considered a controller. The controllermay be implemented by a graphics processing unit (GPU) in addition to, or instead of the CPU.

20 20 20 20 Note that, in some embodiments, the base stationmay include a set of a plurality of physical or logical apparatuses. For example, the base stationaccording to the present embodiment may include a plurality of different kinds of apparatuses such as a baseband unit (BBU) and a radio unit (RU). The base stationmay be understood as a set of the plurality of different kinds of apparatuses. Further, the base stationmay be one of the BBU and the RU or both of them. The BBU and the RU may be connected to each other through a specified interface such as an enhanced common public radio interface (eCPRI).

20 20 The RU may alternatively be phrased as a remote radio unit (RRU) or a radio dot (RD). The RU may correspond to a gNB distributed unit (gNB-DU) described later. The BBU may correspond to a gNB central unit (gNB-CU) described later. The RU may be an apparatus that is formed integrally with an antenna. The antenna of the base station, that is, for example, the antenna formed integrally with the RU may adopt an advanced antenna system, and may support MIMO such as FD-MIMO, or beamforming. The antenna of the base stationmay include, for example, 64 transmission antenna ports and 64 reception antenna ports.

An antenna that is included in the RU may be an antenna panel including at least one antenna element, and the RU may include at least one antenna panel. The RU may include two types of antenna panels that are an antenna panel of horizontally polarized waves and an antenna panel of vertically polarized waves. The RU may include two types of antenna panels that are an antenna panel of right-handed circularly polarized waves and an antenna panel of left-handed circularly polarized waves. The RU may form an independent beam for each antenna panel and control the formed beam.

20 20 20 The base stationsmay be connected to each other. At least one base stationmay be included in a radio access network (RAN). In this case, the base stationmay be simply referred to as a RAN, a RAN node, an access network (AN), or an AN node. A RAN in the LTE may be called an enhanced universal terrestrial RAN (EUTRAN). A RAN in the NR may be called as an NGRAN. A RAN in the W-CDMA (UMTS) may be called a UTRAN.

20 20 The base stationof the LTE may be referred to as an evolved Node B (eNodeB) or an eNB. In this case, the EUTRAN includes at least one eNodeB (eNB). The base stationof the NR may be referred to as a gNodeB or a gNB. In this case, the NGRAN includes at least one gNB. The EUTRAN may include a gNB (en-gNB) connected to a core network in an LTE communication system (EPS) (EPC). The NGRAN may include an ng-eNB connected to a core network 5GC in a 5G communication system (5GS).

20 20 20 20 20 20 20 When the base stationis, for example, an eNB or a gNB, the base stationmay be referred to as a 3GPP access. When the base stationis a wireless access point (an access point), the base stationmay be referred to as a non-3GPP access. The base stationmay be an optical extension apparatus called a remote radio head (RRH). When the base stationis a gNB, the base stationmay be obtained by combining the gNB-CU and gNB-DU described above, or may be one of the gNB-CU and the gNB-DU.

In order to communicate with UE, the gNB-CU hosts a plurality of higher layers (such as RRC, SDAP, and PDCP) from among access strata. The gNB-DU hosts a plurality of lower layers (such as RLC, MAC, and PHY) from among the access strata. From among messages/information that will be described later, RRC signaling (a quasistatic report) may be generated by the gNB-CU, whereas a MAC CE and DCI (dynamic reports) may be generated by the gNB-DU. Alternatively, from among RRC configurations (quasistatic reports), a portion of the configurations such as IE: cellGroupConfig may be generated by the gNB-DU, whereas the remaining configurations may be generated by the gNB-CU. These configurations may be transmitted and received through an F1 interface described later.

20 20 20 20 20 20 20 20 20 20 The base stationmay be capable of communicating with another base station. When a plurality of base stationsis obtained by combining eNBs or by combining an eNB and an en-gNB, the base stationsof the plurality of base stationsmay be connected to each other using an X2 interface. When the plurality of base stationsis obtained by combining gNBs or by combining a gn-eNB and a gNB, the base stationsof the plurality of base stationsmay be connected to each other using an Xn interface. When the plurality of base stations is obtained by combining a gNB-CU and a gNB-DU, the base stationsof the plurality of base stationsmay be connected to each other by the F1 interface described above. A message/information (such as RRC signaling, a MAC control element (MAC CE), or DCI) described later may be transmitted between a plurality of base stationsthrough, for example, the X2 interface, the Xn interface, or the F1 interface.

20 40 A cell provided by the base stationmay be called a serving cell. The concept of the serving cell includes a primary cell (PCell) and a secondary cell (SCell). When dual connectivity is provided to the terminal apparatus, the PCell provided by a master node (MN), and zero SCells or at least one SCell may be called a master cell group. Examples of the dual connectivity include EUTRA-EUTRA dual connectivity, EUTRA-NR dual connectivity (ENDC), EUTRA-NR dual connectivity with 5GC, NR-EUTRA dual connectivity (NEDC), and NR-NR dual connectivity.

40 The serving cell may include a primary secondary cell or a primary SCG cell (PSCell). When the dual connectivity is provided to the terminal apparatus, a PSCell provided by a secondary node (SN), and zero SCells or at least one SCell may be called a secondary cell group (SCG). Unless a special setting (such as PUCCH on SCell) has been performed, a physical uplink control channel (PUCCH) is transmitted in the PCell and the PSCell, but is not transmitted in the SCell. A radio link failure is detected in the PCell and the PSCell, but is not detected in the SCell (does not have to be detected). As described above, the PCell and the PSCell play a special role among the serving cells, and thus are also called special cells (SpCells).

40 40 40 A single downlink component carrier and a single uplink component carrier may be associated with a single cell. A system bandwidth that corresponds to a single cell may be divided into a plurality of bandwidth parts (BWPs). In this case, at least one BWP may be set for the terminal apparatus, and the terminal apparatusmay use one BWP as an active BWP. Radio resources such as a frequency band, the numerology (subcarrier spacing), and a slot format (a slot configuration) that can be used by the terminal apparatusmay be different for respective cells, for respective component carriers, or for respective BWPs. (Configuration of Relay Station)

30 20 30 30 30 The relay stationis a wireless communication apparatus that serves as a relay device for the base station. The relay stationis a type of base station. The relay stationis a type of information processing apparatus. The relay station can alternatively be phrased as a relay base station. The relay stationmay be capable of

40 30 20 40 30 30 20 30 30 20 30 performing the NOMA communication with the terminal apparatus. The relay stationrelays communication between the base stationand the terminal apparatus. The relay stationmay be capable of wirelessly communicating with another relay stationand the base station. The relay stationmay be a ground station apparatus or a non-ground station apparatus. The relay stationforms a radio access network RAN together with the base station. The relay stationmay be a fixed apparatus,

30 30 a movable apparatus, or a floatable apparatus. The size of the coverage of the relay stationis not limited to a specific size. A cell covered by the relay stationmay be a macrocell, a microcell, or a small cell.

30 30 30 The relay stationmay be included in any apparatus in which the relay function of the relay stationis satisfied. The relay stationmay be included in a terminal apparatus such as a smartphone, may be included in, for example, an automobile, a train, or a human-powered vehicle, may be included in, for example, a balloon, an airplane, or a drone, or may be included in a home appliance such as a television, a game machine, an air conditioner, a refrigerator, or a lighting fixture.

30 20 20 30 30 30 The relay stationmay have a configuration similar to the configuration of the base stationdescribed above. As in the case of the base stationdescribed above, the relay stationmay be an apparatus installed in a mobile object, or may be the mobile object itself. As described above, the mobile object may be a mobile terminal such as a smartphone or a mobile phone. The mobile object may be a mobile object that moves on land (on the ground in a narrow sense), or a mobile object that moves under the ground. The mobile object may be a mobile object that moves on water, or a mobile object that moves in water. The mobile object may be a mobile object that moves in the atmosphere of the earth, or a mobile object that moves outside of the atmosphere of the earth. The relay stationmay be a ground station apparatus or a non-ground station apparatus. The relay stationmay be, for example, an aircraft station or a satellite station.

20 30 30 30 30 As in the case of the base station, the coverage of the relay stationmay be as large as a macrocell, or may be as small as a picocell. The coverage of the relay stationmay be as extremely small as a femtocell. The relay stationmay include a beamforming function. For the relay station, a cell or a service area may be formed for each beam.

4 FIG. 4 FIG. 4 FIG. 30 30 31 32 33 35 30 illustrates a configuration of the relay stationaccording to the first embodiment. The relay stationincludes a wireless communication section, a storage, a decoder, and a controller. Note thatillustrates a functional configuration, and a hardware configuration may be different from the configuration illustrated in. Further, functions of the relay stationmay be distributed to be provided to a plurality of physically separated configurations.

31 20 40 30 31 31 31 The wireless communication sectionis a signal processor used to wirelessly communicate with another wireless communication apparatus (such as the base station, the terminal apparatus, or another relay station). The wireless communication sectionsupports at least one radio access scheme. The wireless communication sectionmay support both the NR and the LTE. The wireless communication sectionmay support W-CDMA and cdma3000 in addition to the NR and the LTE.

31 311 312 313 31 311 312 313 31 31 311 312 311 312 313 211 212 213 20 21 20 31 The wireless communication sectionincludes a transmitter, a receiver, and an antenna. The wireless communication sectionmay include a plurality of transmitters, a plurality of receivers, and a plurality of antennas. When the wireless communication sectionsupports a plurality of radio access schemes, the respective components of the wireless communication sectionmay be individually configured for each radio access scheme. The transmitterand the receivermay be individually configured for each of the LTE and the NR. The transmitter, the receiver, and the antennarespectively have configurations similar to the respective configurations of the transmitter, the receiver, and the antennaof the base stationdescribed above. As in the case of the wireless communication sectionof the base station, the wireless communication sectionmay include a beamforming function.

32 32 30 The storageis a data readable/writable storage apparatus such as a DRAM, an SRAM, a flash memory, or a hard disk. The storageserves as a storage device of the relay station.

33 The decoderincludes a function of performing signal processing using an AI/ML model. The AI/ML model is a neural network model obtained by machine learning or deep learning. Examples of the neural network model may include a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, and a long short-term memory (LSTM) model. The AL/ML model may be any of these models, or may be obtained by combining these models in series or in parallel.

33 40 33 23 20 The decoderrestores and outputs first data by performing signal processing using an AI/ML model that uses second data as input, where the second data is received from the terminal apparatusdescribed later. The decodermay have a configuration and a function that are respectively similar to the configuration and the function of the decoderof the base stationdescribed above.

34 30 34 34 30 34 34 The controlleris a controller that controls each component of the relay station. The controllermay be implemented by a processor such as a CPU or an MPU. In particular, the controllermay be implemented by a processor executing, using a RAM or the like as a working region, various programs stored in a storage apparatus situated inside of the relay station. The controllermay be implemented by an integrated circuit such as an ASIC or an FPGA. Any of the CPU, the MPU, the ASIC, and the FPGA can be considered a controller. The controllermay be implemented by a GPU in addition to, or instead of the CPU.

30 30 40 20 Note that the relay stationmay be an IAB relay node. The relay stationoperates as an IAB mobile termination (IAB-MT) for an IAB donor node to which a backhaul is provided, and operates as an IAB distributed unit (IAB-DU) for the terminal apparatusto which access is provided. The IAB donor node may be, for example, the base station, and operates as an IAB central unit (IAB-CU).

40 20 30 40 40 40 40 40 The terminal apparatusis a wireless communication apparatus that wirelessly communicates with another wireless communication apparatus (such as the base station, the relay station, or another terminal apparatus). Examples of the terminal apparatusmay include a mobile phone, a smart device (a smartphone or a tablet), a personal digital assistant (PDA), and a personal computer. The terminal apparatusmay be an apparatus such as a camera for business use that includes a communication function. The terminal apparatusmay be, for example, a motorcycle or mobile relay vehicle that includes a communication apparatus such as a field pickup unit (FPU). The terminal apparatusmay be, for example, a machine-to-machine (M2M) device or an Internet-of-Things (IoT) device.

40 20 40 20 40 40 40 40 40 40 20 40 40 The terminal apparatusmay be capable of performing the NOMA communication with the base station. The terminal apparatusmay be capable of using the automatic retransmission technology such as HARQ upon communicating with the base station. The terminal apparatusmay be capable of performing sidelink communication with another terminal apparatus. The terminal apparatusmay be capable of using the automatic retransmission technology such as HARQ upon performing the sidelink communication. The terminal apparatusmay be capable of performing the NOMA communication upon the sidelink communication with another terminal apparatus. The terminal apparatusmay be capable of performing LPWA communication with another wireless communication apparatus such as the base station. Wireless communication used by the terminal apparatusmay be wireless communication using millimeter waves. The wireless communication used by the terminal apparatusmay be wireless communication using radio waves that includes the sidelink communication, or wireless communication using infrared light or visible light, that is, optical wireless communication.

40 40 40 The terminal apparatusmay be a movable wireless communication apparatus, that is, a mobile object apparatus. The terminal apparatusmay be a wireless communication apparatus installed in a mobile object, or may be the mobile object itself. The terminal apparatusmay be a vehicle, such as an automobile, a bus, a truck, or a motorcycle, that moves on a road, or may be a wireless communication apparatus that is included in the vehicle. The mobile object may be a mobile terminal, or a mobile object that moves on land (on the ground in a narrow sense), under the ground, on water, or in water. The mobile object may be a mobile object, such as a drone or a helicopter, that moves in the atmosphere of the earth, or may be a mobile object, such as an artificial satellite, that moves outside of the atmosphere of the earth.

40 20 20 20 40 40 20 20 The terminal apparatusmay be capable of performing communication by being simultaneously connected to a plurality of base stationsor a plurality of cells. When a single base stationsupports a communication area through a plurality of cells (such as pCell and sCell), the base stationand the terminal apparatuscan communicate with each other by bringing the plurality of cells together using, for example, carrier aggregation (CA), dual connectivity (DC), or multi-connectivity (MC). Alternatively, the terminal apparatuscan also communicate with the base stationsdifferent from each other through cells of the different base stationsusing coordinated multi-point transmission and reception (COMP).

5 FIG. 5 FIG. 5 FIG. 40 40 41 42 43 45 40 illustrates a configuration of the terminal apparatusaccording to the first embodiment. The terminal apparatusincludes a wireless communication section, a storage, an encoder, and a controller. Note thatillustrates a functional configuration, and a hardware configuration may be different from the configuration illustrated in. Further, functions of the terminal apparatusmay be distributed to be provided to a plurality of physically separated configurations.

41 20 30 40 41 45 41 411 412 413 41 411 412 413 21 211 212 213 20 21 20 41 The wireless communication sectionis a signal processor used to wirelessly communicate with another wireless communication apparatus (such as the base station, the relay station, or another terminal apparatus). The wireless communication sectionis controlled by the controller. The wireless communication sectionincludes a transmitter, a receiver, and an antenna. The wireless communication section, the transmitter, the receiver, and the antennarespectively have configurations similar to the respective configurations of the wireless communication section, the transmitter, the receiver, and the antennaof the base station. As in the case of the wireless communication sectionof the base station, the wireless communication sectionmay include a beamforming function.

42 42 40 The storageis a data readable/writable storage apparatus such as a DRAM, an SRAM, a flash memory, or a hard disk. The storageserves as a storage device of the terminal apparatus.

43 The encoderincludes a function of performing signal processing using an AI/ML model. The AI/ML model is a neural network model obtained by machine learning or deep learning. Examples of the neural network model may include a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, and a long short-term memory (LSTM) model. The AL/ML model may be any of these models, or may be obtained by combining these models in series or in parallel.

43 43 43 43 6 FIG. The encodergenerates and outputs second data by performing signal processing using an AI/ML model, where given first data is used as input. The encodermay be implemented by a processor such as a CPU or an MPU. The encodermay be implemented by an integrated circuit such as an ASIC or an FPGA. The configuration of the encoderwill be described in further detail later with reference to.

45 40 45 45 40 45 45 The controlleris a controller that controls each component of the terminal apparatus. The controllermay be implemented by a processor such as a CPU or an MPU. In particular, the controllermay be implemented by a processor executing, using a RAM or the like as a working region, various programs stored in a storage apparatus situated inside of the terminal apparatus. The controllermay be implemented by an integrated circuit such as an ASIC or an FPGA. Any of the CPU, the MPU, the ASIC, and the FPGA can be considered a controller. The controllermay be implemented by a GPU in addition to, or instead of the CPU.

40 43 20 23 40 20 The technology according to the present disclosure is described below on the basis of an example in which AI/ML-model-based signal processing is performed between the terminal apparatusincluding the encoderserving as a signal processor including an AI/ML model, and the base stationincluding the decoderserving as a signal processor including an AI/ML model. However, the terminal apparatusand the base stationmay be replaced with each other. In other words, the technology according to the present disclosure can be applied when signal processing is performed between any two communication apparatuses including AI/ML models.

In the present disclosure, the terms “encode” and “decode” have concepts respectively including any signal processes in a pair, and each of the terms may represent not only signal processing related to compression and decompression of data but also other signal processing. In the present disclosure, the terms “encode” and “decode” may be referred to by terms other than “encode” and “decode”. The technology according to the present disclosure is described below using the terms “encode”, “encoder, “decode”, and “decoder”. However, the technology according to the present disclosure may be described using terms other than the terms described above.

6 FIG. 1 40 20 As illustrated in, the wireless communication systemaccording to the first embodiment includes the terminal apparatusthe base station.

40 43 43 43 The terminal apparatusincludes the encoderserving as a signal processor including an AI/ML model. The encoderencodes given first data using the AI/ML model, where the first data is used as input. Accordingly, the encodergenerates second data.

43 411 20 212 20 The second data generated by the encoderis transmitted by the transmitterto the base stationvia an uplink, and received by the receiverof the base station.

20 23 23 40 23 The base stationincludes the decoderserving as a signal processor including an AI/ML model. The decoderdecodes, using the AI/ML model, the second data received from the terminal apparatus, where the second data is used as input. Accordingly, the decoderrestores the first data.

20 40 40 43 40 20 40 43 In the first embodiment, the base stationtransmits information on which to base learning of an AI/ML model to the terminal apparatusbefore signal processing that corresponds to the above-described encoding based on an AI/ML model, and signal processing that corresponds to the above-described decoding based on an AI/ML model are performed. The terminal apparatussets an AI/ML model included in the encoderof the terminal apparatus, on the basis of information on which to base learning of an AI/ML model, the information being received from the base station. Thereafter, the terminal apparatustrains the AI/ML model included in the encoder.

For example, the information on which to base learning of an AI/ML model may be included in system information, radio-resource-control (RRC) signaling, a MAC control element (MAC CE), or downlink control information (DCI), and the system information, the RRC signaling, the MAC CE, or the DCI may be transmitted. Note that, in the following description, the term “transmission” has a concept including “report” or “configuration”.

Examples of the information on which to base learning of an AI/ML model include pieces of information regarding the type of AI/ML model (such as a CNN model, an RNN model, and an LSTM model), the number of input nodes (the number of neurons of an input layer), the number of output nodes (the number of neurons of an output layer), a layer configuration (such as the number of intermediate layers, the number of neurons of each intermediate layer, and a connection relationship), a value of weighting coefficient (a value of weighting coefficient between neurons), a layer in which learning is performed (such as only performing learning in a fully-connected layer in a final stage), and learning data (such as training data, validation data, and test data).

20 The information on which to base learning may also alternatively be phrased as, for example, an “initial value of learning”. When information on which to base learning of an AI/ML model is defined in advance, this makes it possible to reduce a difference in performance between models that is caused due to learning subsequently performed in respective terminal apparatuses. Thus, the base stationcan perform signal processing together with each terminal apparatus without being conscious of a difference between models of the respective terminal apparatuses.

20 When, for example, the number of input nodes of an AI/ML model and the number of output nodes of the AI/ML model are determined in advance as pieces of information on which to base learning of an AI/ML model, this makes it possible to unify the numbers of input nodes of AI/ML models of respective terminal apparatuses and the numbers of output nodes of the AI/ML models. This results in preventing AI/ML models of respective terminal apparatuses from having different numbers of input nodes and different numbers of output nodes. Thus, the base stationcan transmit data used for signal processing to and receive the data from each terminal apparatus without being conscious of a difference between the respective terminal apparatuses in the numbers of input nodes and output nodes of an AI/ML model.

When, for example, a value of weighting coefficient (an initial value) for an AI/ML model is determined in advance as information on which to base learning of an AI/ML model, this makes it possible to reduce a difference in performance between models that is caused due to learning subsequently performed in respective terminal apparatuses. This results in being able to secure a minimum performance for a terminal apparatus having a low learning capacity.

42 40 42 40 20 40 A portion of information on which to base learning of an AI/ML model may be determined as specifications in advance and stored in the storageof the terminal apparatus. For example, the type of AI/ML model, the number of input nodes, the number of output nodes, and a layer configuration may be determined as specifications in advance and stored in the storageof the terminal apparatus, and information on which to base learning that is transmitted from the base stationto the terminal apparatusmay only include, for example, a value of weighting coefficient and learning data.

40 The information on which to base learning of an AI/ML model may be determined on the basis of a frequency band used by the terminal apparatus. Examples of the frequency band include an uplink band, a downlink band, a sidelink band, a supplementary uplink band, a component carrier, a bandwidth part (BWP), a resource block, a subcarrier, a band of 6 GHZ or less in FR1, a band of from 24 GHz to 54.26 GHz in FR2, a millimeter-wave band, and a tera-wave band.

40 20 20 40 40 20 20 40 When, for example, the terminal apparatusand the base stationcommunicate with each other using a frequency band A, the base stationmay transmit information A on which to base learning to the terminal apparatus, the information A being determined to be adapted for the frequency band A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using a frequency band B, the base stationmay transmit information B on which to base learning to the terminal apparatus, the information B being determined to be adapted for the frequency band B.

40 40 20 40 40 20 40 The information on which to base learning of an AI/ML model may be determined on the basis of a cell in which the terminal apparatusis situated. When, for example, the terminal apparatusis situated in a cell A, the base stationmay transmit information A on which to base learning to the terminal apparatus, the information A being determined to be adapted for communication in the cell A. Further, when, for example, the terminal apparatusis situated in a cell B, the base stationmay transmit information B on which to base learning to the terminal apparatus, the information B being determined to be adapted for communication in the cell B.

40 40 20 40 40 20 40 The information on which to base learning of an AI/ML model may be determined on the basis of a beam used by the terminal apparatus. When, for example, the terminal apparatususes a beam A, the base stationmay transmit information A on which to base learning to the terminal apparatus, the information A being determined to be adapted for the beam A. Further, when, for example, the terminal apparatususes a beam B, the base stationmay transmit information B on which to base learning to the terminal apparatus, the information B being determined to be adapted for the beam B.

40 20 40 20 40 40 20 20 40 The information on which to base learning of an AI/ML model may be determined on the basis of a synchronization signal block (SSB) used by the terminal apparatus. When, for example, the base stationand the terminal apparatuscommunicate with each other using an SSB A, the base stationmay transmit information A on which to base learning to the terminal apparatus, the information A being determined to be adapted for the SSB A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using an SSB B, the base stationmay transmit information B on which to base learning to the terminal apparatus, the information B being determined to be adapted for the SSB B.

40 20 40 20 40 40 20 20 40 The information on which to base learning of an AI/ML model may be determined on the basis of a polarized wave used by the terminal apparatus. When, for example, the base stationand the terminal apparatuscommunicate with each other using a right-handed circularly polarized wave, the base stationmay transmit information R on which to base learning to the terminal apparatus, the information R being determined to be adapted for the right-handed circularly polarized wave. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using a left-handed circularly polarized wave, the base stationmay transmit information L on which to base learning to the terminal apparatus, the information L being determined to be adapted for the left-handed circularly polarized wave.

40 20 40 20 40 40 20 20 40 The information on which to base learning of an AI/ML model may be determined on the basis of a network slice identifier (single network slice selection assistance information (S-NSSAI)) used by the terminal apparatus. When, for example, the base stationand the terminal apparatuscommunicate with each other using a network slice A, the base stationmay transmit information A on which to base learning to the terminal apparatus, the information A being determined to be adapted for the network slice A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using a network slice B, the base stationmay transmit information B on which to base learning to the terminal apparatus, the information B being determined to be adapted for the network slice B.

20 20 40 20 40 20 20 40 20 In the case described above, the base stationmay perform setting, for each network slice, on whether AI/ML-model-based signal processing is to be performed between the base stationand the terminal apparatus. When, for example, the base stationand the terminal apparatuscommunicate with each other using a network slice C, the base stationmay perform setting such that AI/ML-model-based signal processing is performed. Further, when, for example, the base stationand the terminal apparatuscommunicate with each other using a network slice D, the base stationmay perform setting such that the AI/ML-model-based signal processing is not performed.

20 40 40 20 40 40 43 40 40 The base stationmay transmit information on which to base learning of an AI/ML model according to a radio-resource-control state (an RRC state) of the terminal apparatus. When, for example, the terminal apparatusis in an RRC idle state or in an RRC inactive state, the base stationdoes not necessarily have to transmit the information on which to base learning of an AI/ML model to the terminal apparatus. In this case, the terminal apparatusmay individually set and train an AI/ML model included in the encoderof the terminal apparatus, depending on the implementation of the terminal apparatus.

20 40 The base stationmay transmit the information on which to base learning of an AI/ML model to the terminal apparatusat the timing of updating a tracking area (TA), a registration area (RA), or a RAN-based notification area (RNA).

40 20 40 40 43 40 When, for example, the terminal apparatusis in the RRC idle state or in the RRC inactive state, the base stationmay include, in system information, information on which to base learning of an AI/ML model, and may transmit the system information to the terminal apparatus. In this case, the terminal apparatusmay set and train an AI/ML model included in the encoderof the terminal apparatus, on the basis of the received information on which to base learning.

40 20 40 40 43 40 When, for example, the terminal apparatusis in an RRC connected state, the base stationmay include, in RRC signaling, a MAC CE, DCI, or the like, information on which to base learning of an AI/ML model, and may transmit the RRC signaling, the MAC CE, the DCI, or the like to the terminal apparatus. In this case, the terminal apparatusmay set and train an AI/ML model included in the encoderof the terminal apparatus, on the basis of the received information on which to base learning.

The information on which to base learning of an AI/ML model may only include a single piece of information on which to base learning, or may include a plurality of pieces of information on which to base learning.

20 40 43 40 When, for example, information on which to base learning of an AI/ML model that is received from the base stationonly includes a single piece of information on which to base learning, the terminal apparatusmay set and train an AI/ML model included in the encoderof the terminal apparatus, on the basis of the single piece of information on which to base learning.

20 40 40 43 40 When, for example, information on which to base learning of an AI/ML model that is received from the base stationincludes a plurality of pieces of information on which to base learning, the terminal apparatusmay select one of the plurality of pieces of information on which to base learning, and then the terminal apparatusmay set and train an AI/ML model included in the encoderof the terminal apparatus.

20 When information on which to base learning of an AI/ML model is included in system information transmitted from the base stationto each terminal apparatus, the information on which to base learning of an AI/ML model may be shared by cells. Only when each terminal apparatus has the capability of performing AI/ML-model-based signal processing, the terminal apparatus may acquire the information on which to base learning of an AI/ML model, the information being included in the system information.

40 1 40 For example, it is assumed that system information used for an AI/ML model is defined. The terminal apparatusrefers to scheduling information regarding other system information, the scheduling information being included in, for example, a blockof received system information. When the system information used for an AI/ML model is scheduled, the terminal apparatusmay acquire information on which to base learning of an AI/ML model that is included in the system information used for an AI/ML model.

40 43 40 20 The information on which to base learning of an AI/ML model may include information regarding whether additional learning is allowed to be performed. When the information regarding whether additional learning is allowed to be performed is set to “unallowed”, the terminal apparatusis not allowed to additionally train an AI/ML model included in the encoderof the terminal apparatusafter the AI/ML model is set on the basis of the information on which to base learning of an AI/ML model, the information on which to base learning being received from the base station.

20 40 40 43 40 20 Alternatively, separately from the information on which to base learning of an AI/ML model, the base stationmay transmit, to the terminal apparatus, the information regarding whether additional learning is allowed to be performed. When the information regarding whether additional learning is allowed to be performed is set to “unallowed”, the terminal apparatusis not allowed to additionally train an AI/ML model included in the encoderof the terminal apparatusafter the AI/ML model is set on the basis of the information on which to base learning of an AI/ML model, the information on which to base learning being received from the base station.

40 43 40 20 40 20 In the first embodiment, after the terminal apparatussets and trains an AI/ML model included in the encoderof the terminal apparatus, on the basis of information on which to base learning of an AI/ML model, the information being received from the base station, the terminal apparatusmay feed back information regarding the AI/ML model having performed learning (such as the type of AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient) to the base station.

20 40 40 40 43 40 20 The base stationmay transmit, to the terminal apparatus, information (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient) that is used for additional learning of the AI/ML model, on the basis of the information regarding the AI/ML model having performed learning, the information being fed back by the terminal apparatus. The terminal apparatusmay additionally train the AI/ML model included in encoderof the terminal apparatus, on the basis of the information used for additional learning of the AI/ML model, the information being received from the base station.

43 40 40 20 20 When the AI/ML model included in encoderof the terminal apparatushas already performed learning, the terminal apparatusmay transmit information regarding an AI/ML model having performed learning to the base stationbefore receiving, from the base station, information on which to base learning of an AI/ML model, where the information regarding an AI/ML model having performed learning is similar to the above-described information regarding the AI/ML model having performed learning.

40 40 40 When, for example, the terminal apparatusperforms AI/ML-model-based signal processing together with a plurality of base stations, it is necessary for the terminal apparatusto separately train AI/ML models corresponding to the respective base stations. This results in the terminal apparatusbeing very heavily loaded.

43 40 40 20 40 When an AI/ML model included in the encoderof the terminal apparatushas already performed learning, the terminal apparatusmay transmit information regarding its own AI/ML model having performed learning (such as the type of AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient) to the base stationto which the terminal apparatusis newly connected, in order to reduce the load described above.

20 40 20 40 When the base stationreceives, from the terminal apparatus, the information regarding the AI/ML model having performed learning, the base stationmay determine information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient), on the basis of the information regarding the AI/ML model having performed learning, and may transmit the determined information to the terminal apparatus.

40 40 20 40 40 40 20 As capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit, to the base station, the type of AI/ML model supported by the terminal apparatus, the number of input nodes, the number of output nodes, a layer configuration, and the like. Alternatively, as capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit information regarding its own AI/ML model having performed learning to the base station.

20 20 40 40 40 40 40 20 The base stationmay make in advance a list of AI/ML models supported by the base station, and may transmit the list to the terminal apparatusas static information such as the specifications, or may include the list in system information to transmit the system information to the terminal apparatus. The terminal apparatusmay select, from the list, an AI/ML model that can be supported by the terminal apparatus, and may include the selected AI/ML model in capability information regarding the capability of the terminal apparatusto transmit the capability information to the base station.

40 40 20 40 20 40 20 40 For example, as capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit, to the base station, the type of AI/ML model supported by the terminal apparatus, the number of input nodes, the number of output nodes, layer configuration, and the like. When the base stationreceives the capability information from the terminal apparatus, the base stationmay determine information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient), on the basis of the capability information, and may transmit the determined information to the terminal apparatus.

40 40 40 20 In this case, when the terminal apparatushas already caused the AI/ML model to perform learning together with another base station, the terminal apparatusmay transmit limited capability information such as the number of input nodes, the number of output nodes, and a layer configuration with respect to the AI/ML model having performed learning. Alternatively, the terminal apparatusmay transmit information regarding the AI/ML model having already performed learning to the base stationseparately from the capability information.

40 40 20 20 40 In this case, as capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit, to the base station, information regarding an AI/ML model that has performed learning or information regarding an AI/ML model that has not performed learning yet. The base stationmay determine information on which to base learning of an AI/ML model, on the basis of the capability information, and may transmit the determined information to the terminal apparatus.

40 40 20 20 40 For example, as capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit, to the base station, information regarding, for example, a supported AI/ML model, the capability of a CPU or a GPU, whether it is an apparatus that supports reduced capability (RedCap), whether it is an IoT terminal of, for example, NB-IoT or eMTC, and a supported bandwidth. The base stationmay determine information on which to base learning of an AI/ML model, on the basis of the capability information, and may transmit the determined information to the terminal apparatus.

40 40 20 20 20 40 40 As capability information regarding the capability of the terminal apparatus, the terminal apparatusmay transmit information regarding, for example, the remaining processing capability and a maximum value of the processing capability to the base station. For example, such information may be included in capability information upon initial access, RRC signaling, a MAC CE, DCI, or the like, and the capability information, the RRC signaling, the MAC CE, the DCI, or the like may be transmitted. Further, such information may be transmitted when the base stationmakes a request for the information, or may be transmitted regularly. When the information is transmitted regularly, the base stationmay transmit, to the terminal apparatus, an instruction on, for example, how frequent the transmission is to be performed at which timing. Alternatively, the terminal apparatusmay perform transmission with, for example, predetermined frequency and a predetermined timing.

40 43 40 40 40 When the terminal apparatussets and trains a plurality of AI/ML models in the encoderof the terminal apparatuson the basis of information on which to base learning, the information being determined on the basis of a frequency band, the terminal apparatusmay select one of the plurality of AI/ML models according to a frequency band used by the terminal apparatus, and may use the selected AI/ML model.

40 20 40 40 20 40 When, for example, the terminal apparatusand the base stationcommunicate with each other using the frequency band A, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the frequency band A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using the frequency band B, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the frequency band B.

40 43 40 40 40 When the terminal apparatussets and trains a plurality of AI/ML models in the encoderof the terminal apparatuson the basis of information on which to base learning, the information being determined on the basis of a cell, the terminal apparatusmay select one of the plurality of AI/ML models according to a cell in which the terminal apparatusis situated, and may use the selected AI/ML model.

40 40 40 40 When, for example, the terminal apparatusis situated in the cell A, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the cell A. Further, when, for example, the terminal apparatusis situated in the cell B, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the cell B.

40 43 40 40 40 When the terminal apparatussets and trains a plurality of AI/ML models in the encoderof the terminal apparatuson the basis of information on which to base learning, the information being determined on the basis of a beam, the terminal apparatusmay select one of the plurality of AI/ML models according to a beam used by the terminal apparatus, and may use the selected AI/ML model.

40 20 40 40 20 40 When, for example, the terminal apparatusand the base stationcommunicate with each other using the beam A, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the beam A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using the beam B, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the beam B.

40 43 40 40 40 When the terminal apparatussets and trains a plurality of AI/ML models in the encoderof the terminal apparatuson the basis of information on which to base learning, the information being determined on the basis of an SSB, the terminal apparatusmay select one of the plurality of AI/ML models according to an SSB used by the terminal apparatus, and may use the selected AI/ML model.

40 20 40 40 20 40 When, for example, the terminal apparatusand the base stationcommunicate with each other using the SSB_A, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the SSB_A. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using the SSB B, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the SSB_B.

40 43 40 40 40 When the terminal apparatussets and trains a plurality of AI/ML models in the encoderof the terminal apparatuson the basis of information on which to base learning, the information being determined on the basis of a polarized wave, the terminal apparatusmay select one of the plurality of AI/ML models according to a polarized wave used by the terminal apparatus, and may use the selected AI/ML model.

40 20 40 40 20 40 When, for example, the terminal apparatusand the base stationcommunicate with each other using a right-handed circularly polarized wave, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the right-handed circularly polarized wave. Further, when, for example, the terminal apparatusand the base stationcommunicate with each other using a left-handed circularly polarized wave, the terminal apparatusmay use an AI/ML model that is set and trained on the basis of information on which to base learning, the information being determined to be adapted for the left-handed circularly polarized wave.

40 40 The terminal apparatusmay select one of the plurality of AI/ML models according to an RRC state of the terminal apparatus, and may use the selected AI/ML model.

40 40 40 40 40 20 40 For example, when the terminal apparatusis in the RRC idle state, the terminal apparatusmay use an AI/ML model that is individually set and trained depending on the implementation of the terminal apparatus. Alternatively, when the terminal apparatusis in the RRC idle state, the terminal apparatusmay use an AI/ML model that is set on the basis of information on which to base learning, the information being included in system information received from the base station, or the terminal apparatusmay use an AI/ML model that is set and trained on the basis of the information on which to base learning.

40 40 40 40 40 20 40 For example, when the terminal apparatusis in the RRC inactive state, the terminal apparatusmay use an AI/ML model that is set on the basis of information on which to base learning, the information being received during an RRC connection, or the terminal apparatusmay use an AI/ML model that is set and trained on the basis of the information on which to base learning. Alternatively, when the terminal apparatusis in the RRC inactive state, the terminal apparatusmay use an AI/ML model that is set on the basis of information on which to base learning, the information being included in system information received from the base station, or the terminal apparatusmay use an AI/ML model that is set and trained on the basis of the information on which to base learning.

40 40 40 When, for example, the terminal apparatusis in the RRC connected state, the terminal apparatusmay use an AI/ML model that is set on the basis of information on which to base learning, the information being included in, for example, RRC signaling, a MAC CE, or DCI, or the terminal apparatusmay use an AI/ML model that is set and trained on the basis of the information on which to base learning.

7 FIG. 50 50 50 illustrates a configuration of a data set providing apparatusaccording to the first embodiment. The data set providing apparatusis an apparatus that provides information on which to base learning of an AI/ML model, and can be accessed in common by terminal apparatuses of different vendors and by base stations of different vendors. The data set providing apparatusmay be an apparatus that has a physical body, or may be defined as, for example, a single network function (NF).

50 51 52 53 54 51 52 20 40 53 20 40 54 51 52 53 The data set providing apparatusincludes a storage, a receiver, a transmitter, and a controller. The storagestores therein information on which to base learning of an AI/ML model. The receiverreceives, from the base stationor the terminal apparatus, a request for provision of information on which to base learning of an AI/ML model. The transmittertransmits information on which to base learning of an AI/ML model to the base stationor terminal apparatushaving transmitted the request for provision. The controllercontrols the storage, the receiver, and the transmitter.

20 40 50 20 40 50 The base stationor the terminal apparatusaccesses an entity of the data set providing apparatusbefore training of an AI/ML model, and receives in advance information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, a value of weighting coefficient, a layer in which learning is performed, and learning data). This enables the base stationand the terminal apparatusto acquire, from the data set providing apparatusin common, one piece of information on which to base learning, the piece of information not depending on the specifications or the implementation of a specific vendor. This makes it possible to set and train an AI/ML model that is interoperable among different vendors.

50 The data set providing apparatusmay be defined as a first data set providing apparatus intended for base station, and a second data set providing apparatus intended for terminal apparatus, the first and second data set providing apparatuses being defined separately from each other. Alternatively, only one of the first data set providing apparatus intended for base station and the second data set providing apparatus intended for terminal apparatus may be defined.

20 40 When, for example, two data set providing apparatuses are defined separately from each other, the base stationreceives in advance information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, a value of weighting coefficient, a layer in which learning is performed, and learning data) from the first data set providing apparatus. Further, the terminal apparatusreceives in advance information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, a value of weighting coefficient, a layer in which learning is performed, and learning data) from the second data set providing apparatus.

20 40 50 50 50 20 40 The base stationor the terminal apparatusmay transmit, to the data set providing apparatus, a request for provision of information on which to base learning of an AI/ML model. When the data set providing apparatusreceives the request for provision, the data set providing apparatusmay newly transmit information on which to base learning of an AI/ML model to the base stationor terminal apparatushaving transmitted the request for provision.

20 40 50 50 20 40 20 40 50 50 As capability information regarding its own capability, the base stationor the terminal apparatusmay transmit, to the data set providing apparatus, information regarding, for example, a currently held AI/ML model that has performed learning, a supported AI/ML model, the capability of a CPU or a GPU, whether it is an apparatus that supports reduced capability (RedCap), whether it is an IoT terminal of, for example, NB-IoT or eMTC, and a supported bandwidth. The data set providing apparatusmay determine information on which to base learning of an AI/ML model on the basis of the capability information, and may transmit the determined information to the base stationor the terminal apparatus. Further, on the basis of the information regarding the currently held AI/ML model having performed learning, the information being received from the base stationor the terminal apparatus, the data set providing apparatusmay update information on which to base learning of an AI/ML model, the updated information being held by the data set providing apparatus.

20 40 50 20 40 40 50 The base stationmay transmit, to the terminal apparatus, information regarding access to the data set providing apparatus. For example, the base stationmay transmit, to the terminal apparatus, information such as an IP address or MAC address that is necessary for the terminal apparatusto access the data set providing apparatus.

8 FIG. 50 50 50 50 20 50 50 50 50 50 50 illustrates an example of a configuration of a wireless communication network that includes the data set providing apparatus. For example, the data set providing apparatusmay be arranged under a data network that is beyond a core network, as illustrated in the figure. For example, the data set providing apparatusmay be included in a server (such as a mobile edge computing (MEC) server) that is connected to a user plane function (UPF). For example, the data set providing apparatusmay be included in a server (such as a MEC server) that is connected to the base station. For example, the data set providing apparatusmay be defined as a new network function (NF). For example, the data set providing apparatusmay be defined as an additional function of an existing NF (such as a model training logical function (MTLF)). Further, not only a single data set providing apparatusarranged at a single position but also a plurality of data set providing apparatusesarranged at a plurality of positions may be adopted, or not only a single data set providing apparatusbut also a plurality of data set providing apparatusesmay be defined.

40 23 When the terminal apparatusis handed over from a certain base station (a source base station) to another base station (a target base station), the source base station may transmit, to the target base station, information regarding an AI/ML model that has performed learning (such as the type of AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient), information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient), and capability information regarding the capability for an AI/ML model (such as the type of a supported AI/ML model, the number of input nodes, the number of output nodes, and a layer configuration), and other information. When the target base station receives these pieces of information from the source base station, the target base station may set and train an AI/ML model included in the decoderof the target base station, on the basis of the received information.

40 40 40 40 40 40 On the basis of, for example, capability information regarding the capability of the terminal apparatusor a base station with respect to an AI/ML model, or information regarding an AI/ML model that has performed learning, a base station to which the terminal apparatusis handed over may be limited or determined. In other words, when the terminal apparatusis handed over, there will be a need to newly set and train an AI/ML model if an AI/ML model that has already performed learning is not allowed to be used at a base station to which the terminal apparatusis handed over. This results in an increase in processing burdens imposed on the terminal apparatusand the base station. Thus, the terminal apparatusmay be handed over to a base station that can use an AI/ML model that has already performed learning.

1. Only the base station uses an AI/ML model. 2. Only the terminal apparatus uses an AI/ML model. 3. Both the terminal apparatus and the base station use AI/ML models (the case of the first embodiment). In general, the following are three use cases in which a terminal apparatus and a base station perform AI/ML-model-based signal processing together.

In the case of “1” described above, only the base station uses an AI/ML model. Thus, it is not a problem if AI/ML models of different base stations are different. When, for example, there is a plurality of base station vendors, each of the plurality of base station vendors can set and train an AI/ML model individually depending on the implementation.

In the case of “2” described above, only the terminal apparatus uses an AI/ML model. Thus, it is not a problem if AI/ML models of different terminal apparatuses are different. When, for example, there is a plurality of terminal vendors, each of the plurality of terminal vendors can set and train an AI/ML model individually depending on the implementation.

However, there may be a difference in processing capability between terminal apparatuses, and there may be a terminal apparatus that is not capable of setting and training an AI/ML model properly. In order to deal with this issue, the base station may transmit, to the terminal apparatus, information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient). Further, the terminal apparatus may transmit, to the base station, information regarding the capability of the terminal apparatus with respect to an AI/ML model (such as the type of a supported AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, and information regarding a held CPU or GPU).

In the case of “3” described above, both the terminal apparatus and the base station use AI/ML models. Thus, there is a need to consider a difference between the terminal apparatus and the base station. In order to deal with this issue, the base station may transmit, to the terminal apparatus, information on which to base learning of an AI/ML model (such as the number of input nodes, the number of output nodes, a layer configuration, and a value of weighting coefficient). Further, the terminal apparatus may transmit, to the base station, information regarding the capability of the terminal apparatus with respect to an AI/ML model (such as the type of a supported AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, and information regarding a held CPU or GPU).

The technology according to the present disclosure can also be applied to communication between terminal apparatuses such as sidelink and communication between base stations such as integrated access backhaul (IAB).

When, for example, there are a parent terminal apparatus and a child terminal apparatus in communication between terminal apparatuses, the parent terminal apparatus may serve a function of the base station described above, and the child terminal apparatus may serve a function of the terminal apparatus described above.

When, for example, there are a parent base station and a child base station in communication between base stations, the parent base station may serve a function of the base station described above, and the child base station may serve a function of the terminal apparatus described above.

Six examples of a procedure in which the technology according to the present disclosure is applied to channel-state-information (CSI) feedback are described below. However, a range in which the technology according to the present disclosure can be applied is not limited thereto. For example, the technology according to the present disclosure can also be applied to, for example, beam management, beam positioning, or mobility management.

9 FIG. is a diagram used to describe an example of a procedure in which a base station includes, in system information, information on which to base learning of an AI/ML model and transmits the system information to each of different terminal apparatuses.

101 In Step S, each of terminal apparatuses A and B receives a synchronization signal transmitted by a base station, and performs downlink synchronization. Further, each of the terminal apparatuses A and B receives system information transmitted by the base station to receive information necessary for cell connection. Here, the system information includes information on which to base learning of an AI/ML model.

102 In Step S, each of the terminal apparatuses A and B may additionally train an AI/ML model included in an encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information received from the base station. The base station may additionally train an AI/ML model included in a decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information transmitted by the base station.

Each of the terminal apparatuses A and B may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform additional training individually.

103 In Step S, each of the terminal apparatuses A and B performs a random access procedure, and establishes connection with the base station. Each of the terminal apparatuses A and B achieves uplink synchronization by the random access procedure, and completes the connection with the base station. The random access procedure may be performed after completion of additional training or during the additional training.

104 105 In Step S, each of the terminal apparatuses A and B transmits capability information regarding the capability of a corresponding one of the terminal apparatuses A and B to the base station. The capability information may include capability information regarding the capability for an AI/ML model. In Step S, the base station transmits quasistatic information to each of the terminal apparatuses A and B. The quasistatic information may be, for example, RRC signaling.

106 107 In Step S, the base station transmits a reference signal for estimating a downlink channel state to the terminal apparatus B. In Step S, the terminal apparatus B performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

108 109 In Step S, the terminal apparatus B performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus B to encode the downlink-channel-state information. In Step S, the terminal apparatus B feeds back the encoded downlink-channel-state information to the base station.

The downlink-channel-state information described above may be an uplink control signal (such as a channel quality indicator (CQI), a precoding matrix indicator (PMI), or a resource indicator (RI)) that is defined by uplink control information (UCI), may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

110 In Step S, the base station performs signal processing based on an AI/ML model included in the decoder of the base station to decode the encoded downlink-channel-state information received from the terminal apparatus B, and acquires the downlink-channel-state information.

111 In Step S, the base station transmits a downlink control signal. The downlink control signal may be, for example, downlink control information (DCI). The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

112 113 In Step S, the base station transmits downlink data. The downlink data may be, for example, a physical downlink shared channel (PDSCH). In Step S, the terminal apparatus B transmits information regarding retransmission control to the base station according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be acknowledge (ACK)/negative acknowledge (NACK), or may be “hybrid automatic repeat request-acknowledge” (HARQ-ACK).

10 FIG. is a diagram used to describe an example of a procedure in which a base station includes, in quasistatic information (RRC signaling), information on which to base learning of an AI/ML model and transmits the quasistatic information to each of different terminal apparatuses.

201 In Step S, each of terminal apparatuses A and B receives a synchronization signal transmitted by a base station, and performs downlink synchronization. Further, each of the terminal apparatuses A and B receives system information transmitted by the base station to receive information necessary for cell connection.

202 In Step S, each of the terminal apparatuses A and B performs a random access procedure, and establishes connection with the base station. Each of the terminal apparatuses A and B achieves uplink synchronization by the random access procedure, and completes the connection with the base station. The random access procedure may be performed after completion of additional training described later or during the additional training.

203 In Step S, each of the terminal apparatuses A and B transmits capability information regarding the capability of a corresponding one of the terminal apparatuses A and B to the base station. The capability information may include capability information regarding the capability for an AI/ML model.

204 In Step S, the base station transmits quasistatic information to each of the terminal apparatuses A and B. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes information on which to base learning of an AI/ML model.

205 In Step S, each of the terminal apparatuses A and B may additionally train an AI/ML model included in an encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information received from the base station. The base station may additionally train an AI/ML model included in a decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information transmitted by the base station.

Each of the terminal apparatuses A and B may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform additional training individually.

206 207 In Step S, the base station transmits a reference signal for estimating a downlink channel state to the terminal apparatus B. In Step S, the terminal apparatus B performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

208 209 In Step S, the terminal apparatus B performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus B to encode the downlink-channel-state information. In Step S, the terminal apparatus B feeds back the encoded downlink-channel-state information to the base station.

The downlink-channel-state information described above may be an uplink control signal (such as a COI, a PMI, or an RI) that is defined by UCI, may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

210 In Step S, the base station performs signal processing based on an AI/ML model included in the decoder of the base station to decode the encoded downlink-channel-state information received from the terminal apparatus B, and acquires the downlink-channel-state information.

211 In Step S, the base station transmits a downlink control signal. The downlink control signal may be, for example, DCI. The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

212 213 In Step S, the base station transmits downlink data. The downlink data may be, for example, a PDSCH. In Step S, the terminal apparatus B transmits information regarding retransmission control to the base station according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be ACK/NACK, or may be HARQ-ACK.

11 FIG. is a diagram used to describe an example of a procedure in which a base station includes, in each of system information and quasistatic information (RRC signaling), information on which to base learning of an AI/ML model and transmits the system information and the quasistatic information to each of different terminal apparatuses.

301 In Step S, each of terminal apparatuses A and B receives a synchronization signal transmitted by a base station, and performs downlink synchronization. Further, each of the terminal apparatuses A and B receives system information transmitted by the base station to receive information necessary for cell connection. Here, the system information includes information on which to base learning of an AI/ML model.

302 In Step S, each of the terminal apparatuses A and B may additionally train an AI/ML model included in an encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information received from the base station. The base station may additionally train an AI/ML model included in a decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information transmitted by the base station.

Each of the terminal apparatuses A and B may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform additional training individually.

303 In Step S, each of the terminal apparatuses A and B performs a random access procedure, and establishes connection with the base station. Each of the terminal apparatuses A and B achieves uplink synchronization by the random access procedure, and completes the connection with the base station. The random access procedure may be performed after completion of additional training or during the additional training.

304 In Step S, each of the terminal apparatuses A and B transmits capability information regarding the capability of a corresponding one of the terminal apparatuses A and B to the base station. The capability information may include capability information regarding the capability for an AI/ML model.

305 In Step S, the base station transmits quasistatic information to each of the terminal apparatuses A and B. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes information on which to base learning of an AI/ML model.

306 In Step S, each of the terminal apparatuses A and B may further additionally train an AI/ML model included in the encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information received from the base station. The base station may further additionally train an AI/ML model included in the decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information transmitted by the base station.

Each of the terminal apparatuses A and B may determine whether to perform further additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform further additional training individually.

307 308 In Step S, the base station transmits a reference signal for estimating a downlink channel state to the terminal apparatus B. In Step S, the terminal apparatus B performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

309 310 In Step S, the terminal apparatus B performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus B to encode the downlink-channel-state information. In Step S, the terminal apparatus B feeds back the encoded downlink-channel-state information to the base station.

The downlink-channel-state information described above may be an uplink control signal (such as a CQI, a PMI, or an RI) that is defined by UCI, may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

311 In Step S, the base station performs signal processing based on an AI/ML model included in the decoder of the base station to decode the encoded downlink-channel-state information received from the terminal apparatus B, and acquires the downlink-channel-state information.

312 In Step S, the base station transmits a downlink control signal. The downlink control signal may be, for example, DCI. The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

313 314 In Step S, the base station transmits downlink data. The downlink data may be, for example, a PDSCH. In Step S, the terminal apparatus B transmits information regarding retransmission control to the base station according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be ACK/NACK, or may be HARQ-ACK.

12 FIG. is a diagram used to describe an example of a procedure in which a base station receives, from a data set providing apparatus, information on which to base learning of an AI/ML model; includes, in each of system information and quasistatic information (RRC signaling), the information on which to base learning of an AI/ML model; and transmits the system information and the quasistatic information to each of different terminal apparatuses.

401 402 In Step S, a base station transmits, to a data set providing apparatus, a request for provision of information on which to base learning of an AI/ML model. In Step S, the data set providing apparatus transmits the information on which to base learning of an AI/ML model to the base station.

403 In Step S, the base station may additionally train an AI/ML model included in a decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being received from the data set providing apparatus.

404 In Step S, each of terminal apparatuses A and B receives a synchronization signal transmitted by the base station, and performs downlink synchronization. Further, each of the terminal apparatuses A and B receives system information transmitted by the base station to receive information necessary for cell connection. Here, the system information includes the information on which to base learning of an AI/ML model.

405 In Step S, each of the terminal apparatuses A and B may additionally train an AI/ML model included in an encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information received from the base station.

Each of the terminal apparatuses A and B may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform additional training individually.

406 In Step S, each of the terminal apparatuses A and B performs a random access procedure, and establishes connection with the base station. Each of the terminal apparatuses A and B achieves uplink synchronization by the random access procedure, and completes the connection with the base station. The random access procedure may be performed after completion of additional training or during the additional training.

407 408 In Step S, each of the terminal apparatuses A and B transmits capability information regarding the capability of a corresponding one of the terminal apparatuses A and B to the base station. The capability information may include capability information regarding the capability for an AI/ML model. In Step S, the base station transmits quasistatic information to each of the terminal apparatuses A and B. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes the information on which to base learning of an AI/ML model.

409 In Step S, each of the terminal apparatuses A and B may further additionally train an AI/ML model included in the encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information received from the base station. The base station may further additionally train an AI/ML model included in the decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information transmitted by the base station.

Each of the terminal apparatuses A and B may determine whether to perform further additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform further additional training individually.

410 411 In Step S, the base station transmits a reference signal for estimating a downlink channel state to the terminal apparatus B. In Step S, the terminal apparatus B performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

412 413 In Step S, the terminal apparatus B performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus B to encode the downlink-channel-state information. In Step S, the terminal apparatus B feeds back the encoded downlink-channel-state information to the base station.

The downlink-channel-state information described above may be an uplink control signal (such as a CQI, a PMI, or an RI) that is defined by UCI, may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

414 In Step S, the base station performs signal processing based on an AI/ML model included in the decoder of the base station to decode the encoded downlink-channel-state information received from the terminal apparatus B, and acquires the downlink-channel-state information.

415 In Step S, the base station transmits a downlink control signal. The downlink control signal may be, for example, DCI. The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

416 417 In Step S, the base station transmits downlink data. The downlink data may be, for example, a PDSCH. In Step S, the terminal apparatus B transmits information regarding retransmission control to the base station according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be ACK/NACK, or may be HARQ-ACK.

13 FIG. is a diagram used to describe an example of a procedure in which each of a base station and a terminal apparatus receives, from a data set providing apparatus, information on which to base learning of an

AI/ML model.

501 502 In Step S, a base station transmits, to a data set providing apparatus, a request for provision of information on which to base learning of an AI/ML model. In Step S, the data set providing apparatus transmits the information on which to base learning of an AI/ML model to the base station.

503 In Step S, the base station may additionally train an AI/ML model included in a decoder of the base station, on the basis of the information on which to base learning of an AI/ML model, the information being received from the data set providing apparatus.

504 In Step S, each of terminal apparatuses A and B receives a synchronization signal transmitted by the base station, and performs downlink synchronization.

Further, each of the terminal apparatuses A and B receives system information transmitted by the base station to receive information necessary for cell connection. Here, the system information includes information regarding the data set providing apparatus.

505 506 In Step S, each of the terminal apparatuses A and B transmits, to the data set providing apparatus, a request for provision of the information on which to base learning of an AI/ML model, on the basis of the information regarding the data set providing apparatus, the information regarding the data set providing apparatus being included in the system information received from the base station. In Step S, the data set providing apparatus transmits the information on which to base learning of an AI/ML model to each of the terminal apparatuses A and B.

507 In Step S, each of the terminal apparatuses A and B may additionally train an AI/ML model included in an encoder of a corresponding one of the terminal apparatuses A and B, on the basis of the information on which to base learning of an AI/ML model, the information being received from the data set providing apparatus.

Each of the terminal apparatuses A and B may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station. The terminal apparatuses A and B may obtain different training results (learning results) since the terminal apparatuses A and B perform additional training individually.

508 In Step S, each of the terminal apparatuses A and B performs a random access procedure, and establishes connection with the base station. Each of the terminal apparatuses A and B achieves uplink synchronization by the random access procedure, and completes the connection with the base station. The random access procedure may be performed after completion of additional training or during the additional training.

509 In Step S, each of the terminal apparatuses A and B transmits capability information regarding the capability of a corresponding one of the terminal apparatuses A and B to the base station. The capability information may include capability information regarding the capability for an AI/ML model.

510 In Step S, the base station transmits quasistatic information to each of the terminal apparatuses A and B. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes information regarding the technology according to the present disclosure.

511 512 In Step S, the base station transmits a reference signal for estimating a downlink channel state to the terminal apparatus B. In Step S, the terminal apparatus B performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

513 In Step S, the terminal apparatus B performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus B to encode the downlink-channel-state information. In Step

514 S, the terminal apparatus B feeds back the encoded downlink-channel-state information to the base station.

The downlink-channel-state information described above may be an uplink control signal (such as a CQI, a PMI, or an RI) that is defined by UCI, may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

515 In Step S, the base station performs signal processing based on an AI/ML model included in the decoder of the base station to decode the encoded downlink-channel-state information received from the terminal apparatus B, and acquires the downlink-channel-state information.

516 In Step S, the base station transmits a downlink control signal. The downlink control signal may be, for example, DCI. The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

517 518 In Step S, the base station transmits downlink data. The downlink data may be, for example, a PDSCH. In Step S, the terminal apparatus B transmits information regarding retransmission control to the base station according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be ACK/NACK, or may be HARQ-ACK.

14 FIG. is a diagram used to describe an example of a procedure in which a terminal apparatus is handed over to a different base station.

601 In Step S, a terminal apparatus receives a synchronization signal transmitted by a base station C, and performs downlink synchronization. Further, the terminal apparatus receives system information transmitted by the base station C to receive information necessary for cell connection. Here, the system information includes information on which to base learning of an AI/ML model.

602 In Step S, the terminal apparatus may additionally train an AI/ML model included in an encoder of the terminal apparatus, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information received from the base station C. The base station C may additionally train an AI/ML model included in a decoder of the base station C, on the basis of the information on which to base learning of an AI/ML model, the information being included in the system information transmitted by the base station C. The terminal apparatus may determine whether to perform additional training, depending on the implementation or on the basis of an instruction transmitted by the base station C.

603 In Step S, the terminal apparatus performs a random access procedure, and establishes connection with the base station C. The terminal apparatus achieves uplink synchronization by the random access procedure, and completes the connection with the base station C. The random access procedure may be performed after completion of additional training or during the additional training.

604 In Step S, the terminal apparatus transmits capability information regarding the capability of the terminal apparatus to the base station C. The capability information may include capability information regarding the capability for an AI/ML model.

605 In Step S, the base station C transmits quasistatic information to the terminal apparatus. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes information on which to base learning of an AI/ML model.

606 In Step S, the terminal apparatus may further additionally train an AI/ML model included in the encoder of the terminal apparatus, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information received from the base station C. The base station C may further additionally train an AI/ML model included in the decoder of the base station C, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information transmitted by the base station C. The terminal apparatus may determine whether to perform further additional training, depending on the implementation or on the basis of an instruction transmitted by the base station c.

607 608 In Step S, the base station C transmits a reference signal for estimating a downlink channel state to the terminal apparatus. In Step S, the terminal apparatus performs processing of estimating a downlink channel state, and generates downlink-channel-state information.

609 610 In Step S, the terminal apparatus performs signal processing based on an AI/ML model included in the encoder of the terminal apparatus to encode the downlink-channel-state information. In Step S, the terminal apparatus feeds back the encoded downlink-channel-state information to the base station C.

The downlink-channel-state information described above may be an uplink control signal (such as a CQI, a PMI, or an RI) that is defined by UCI, may be channel response information itself, or may be eigenvalue information regarding an eigenvalue of a channel.

611 In Step S, the base station C performs signal processing based on an AI/ML model included in the decoder of the base station C to decode the encoded downlink-channel-state information received from the terminal apparatus, and acquires the downlink-channel-state information.

612 In Step S, the base station C transmits a downlink control signal. The downlink control signal may be, for example, DCI. The downlink control signal may include information regarding the technology according to the present disclosure. Further, the downlink control signal may be partially determined on the basis of the downlink-channel-state information fed back above.

613 614 In Step S, the base station C transmits downlink data. The downlink data may be, for example, a PDSCH. In Step S, the terminal apparatus transmits information regarding retransmission control to the base station C according to a result of decoding for the downlink data. For example, the information regarding retransmission control may be ACK/NACK, or may be HARQ-ACK.

615 616 In Step S, the base station C (a source cell) and the terminal apparatus perform a measurement procedure related to a connection cell. In Step S, the base station C determines, on the basis of a result of the measurement procedure, whether there is a need for handover, and determines that there is a need for handover.

617 618 619 In Step S, the base station C transmits a handover request to a base station D that is a candidate target cell. In Step S, the base station D performs admission control. In Step S, the base station D transmits, to the base station C, a report (handover request acknowledge) about whether handover is permitted.

620 621 In Step S, the base station C transmits, to the base station D, information on which to base learning of an AI/ML model held by the base station C. Further, the base station C may transmit information regarding its own AI/ML model having performed learning to the base station D. In Step S, the base station D transmits ACK to the base station C.

622 In Step S, the base station D may further additionally train an AI/ML model included in a decoder of the base station D, on the basis of the information on which to base learning of an AI/ML model, the information being received from the base station C.

623 624 In Step S, the base station C performs RCC configuration on the terminal apparatus, and transmits information regarding handover to the terminal apparatus. The information regarding handover includes information on which to base learning of an AI/ML model. In Step S, the terminal apparatus is detached from the base station C.

625 In Step S, the terminal apparatus performs a random access procedure, and establishes connection with the base station D. The terminal apparatus achieves uplink synchronization by the random access procedure, and completes the connection with the base station D. The random access procedure may be performed after completion of additional training or during the additional training.

626 In Step S, the terminal apparatus transmits capability information regarding the capability of the terminal apparatus to the base station D. The capability information may include capability information regarding the capability for an AI/ML model.

627 In Step S, the base station D transmits quasistatic information to the terminal apparatus. The quasistatic information may be, for example, RRC signaling. The quasistatic information includes information on which to base learning of an AI/ML model.

628 In Step S, the terminal apparatus may further additionally train an AI/ML model included in the encoder of the terminal apparatus, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information received from the base station D. The base station D may further additionally train an AI/ML model included in the decoder of the base station D, on the basis of the information on which to base learning of an AI/ML model, the information being included in the quasistatic information transmitted by the base station D. The terminal apparatus may determine whether to perform further additional training, depending on the implementation or on the basis of an instruction transmitted by the base station D.

40 20 43 23 20 40 40 20 As described above, a communication apparatus (the terminal apparatusor the base station) according to the first embodiment of the present disclosure includes a signal processor (the encoderor the decoder) that includes an AI/ML model, a receiver that receives, from another communication apparatus (the base stationor the terminal apparatus), information on which to base learning, and a controller that sets the AI/ML model on the basis of the information on which to base learning. Such characteristics enable the communication apparatus (the terminal apparatusor the base station) according to the first embodiment of the present disclosure to perform signal processing without being conscious of a difference between AI/ML models.

50 50 51 52 20 40 50 Further, the data set providing apparatusaccording to the first embodiment of the present disclosure is a data set providing apparatus that can be accessed in common by communication apparatuses of different vendors. The data set providing apparatusincludes the storagestoring therein information on which to base learning of an AI/ML model, the receiverreceiving, from a communication apparatus, a request for provision of the information on which to base learning of an AI/ML model, and a transmitter that transmits the information on which to base learning of an AI/ML model to the communication apparatus. Such characteristics enable the base stationand the terminal apparatusto acquire, from the data set providing apparatusin common, one piece of information on which to base learning, the piece of information not depending on the specifications or the implementation of a specific vendor. This makes it possible to set and train an AI/ML model that is interoperable among different vendors.

The technology according to the present disclosure is not limited to a specific standard, and any modifications may be made to the illustrated settings as appropriate. Note that the embodiments described above are examples used to embody the technology according to the present disclosure, and the technology according to the present disclosure may be embodied in a variety of other embodiments. For example, various modifications, replacements, omissions, or combinations thereof may be applied without departing from the scope of the present disclosure. Embodiments obtained by applying, for example, such modifications, replacements, omissions, or combinations are also included in the scope of the present disclosure, as well as the scope of claimed embodiments of the present disclosure and their equivalents.

The processing procedures described in the present disclosure may be considered a method including a series of the procedures. Alternatively, the processing procedures described in the present disclosure may be considered a program used to cause a computer to perform the series of the procedures, or a recording medium that has stored therein the program. Further, the processing described above may be performed by a processor such as a CPU of a computer. Further, the type of the recording medium is not particularly limited since the embodiments of the present disclosure are not affected by the type of the recording medium.

2 7 FIGS.to The structural elements illustrated inin the present disclosure may be provided by software or may be implemented by hardware. For example, each structural element may be a software module provided by software such as a microprogram, and may be provided by a processor executing the software module. Alternatively, each structural element may be implemented by a circuit block on a semiconductor chip (a die), that is, for example, an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). Further, the number of structural elements does not necessarily have to be equal to the number of pieces of hardware by which the structural elements are implemented. For example, a plurality of structural elements may be implemented by a single processor or circuit. Conversely, a single structural element may be implemented by a plurality of processors or circuits.

The type of the processor described in the present disclosure is not limited. For example, the processor may be a CPU, a micro processing unit (MPU), or a graphics processing unit (GPU).

a signal processor that includes an AI/ML model; a receiver that receives, from another communication apparatus, information on which to base learning; and a controller that sets the AI/ML model on the basis of the information on which to base learning. [1] A communication apparatus, including: the information on which to base learning includes at least one of the type of AI/ML model, the number of input nodes, the number of output nodes, a layer configuration, a value of weighting coefficient, a layer in which learning is performed, or learning data. [2] The communication apparatus according to [1], in which the information on which to base learning is determined on the basis of at least one of a cell in which the communication apparatus is situated, a frequency band used by the communication apparatus, or a synchronization signal block (SSB) used by the communication apparatus. [3] The communication apparatus according to [1] or [2], in which the information on which to base learning is included in one of system information, radio-resource-control (RRC) signaling, a MAC control element (MAC CE), and downlink control information (DCI) that are transmitted by the other communication apparatus. [4] The communication apparatus according to any one of [1] to [3], in which the information on which to base learning is included in the system information when the communication apparatus is in a radio-resource-control (RRC) idle state or in a radio-resource-control (RRC) inactive state. [5] The communication apparatus according to [4], in which only when the communication apparatus has the capability of performing AI/ML-model-based signal processing, the controller acquires the information on which to base learning, the information being included in the system information. [6] The communication apparatus according to [5], in which the information on which to base learning is included in the RRC signaling, the MAC CE, or the DCI when the communication apparatus is in a radio-resource-control (RRC) connected state. [7] The communication apparatus according to [4], in which the receiver further receives, from the other communication apparatus, information regarding whether additional learning is allowed to be performed, and when the information regarding whether additional learning is allowed to be performed is set to “unallowed”, the controller does not additionally train the AI/ML model after the AI/ML model is set on the basis of the information on which to base learning. [8] The communication apparatus according to any one of [1] to [7], in which the controller trains the AI/ML model after the AI/ML model is set on the basis of the information on which to base learning. [9] The communication apparatus according to any one of [1] to [7], in which a transmitter that transmits a signal to the other communication apparatus, in which after the training of the AI/ML model is performed, the controller feeds back information regarding the AI/ML model having performed learning to the other communication apparatus through the transmitter. [10] The communication apparatus according to [9], further including a transmitter that transmits a signal to the other communication apparatus, in which when the AI/ML model has already performed learning, the controller transmits information regarding the AI/ML model having performed learning to the other communication apparatus through the transmitter before the information on which to base learning is received from the other communication apparatus. [11] The communication apparatus according to any one of [1] to [10], further including a transmitter that transmits a signal to the other communication apparatus, in which the controller transmits capability information regarding the capability of the communication apparatus to the other communication apparatus through the transmitter. [12] The communication apparatus according to any one of [1] to [11], further including the capability information includes information regarding an AI/ML model supported by the communication apparatus. [13] The communication apparatus according to [12], in which [14] The communication apparatus according to or [13], in which the capability information includes information regarding the AI/ML model being included in the communication apparatus and having performed learning. [15] The communication apparatus according to any one of [12] to [14], in which the capability information includes one of a piece of information regarding the remaining processing capability of the communication apparatus and a piece of information regarding a maximum value of the processing capability of the communication apparatus, or both of the pieces of information. the capability information is transmitted when the other communication apparatus makes a request for the capability information. [16] The communication apparatus according to [15], in which the capability information is transmitted regularly on the basis of an instruction on how frequent the transmission is to be performed at which timing, the instruction being received from the other communication apparatus, or the capability information is transmitted regularly on the basis of predetermined frequency and a predetermined timing. [17] The communication apparatus according to [15], in which the signal processor includes a plurality of the AI/ML models, and the controller selects one of the plurality of the AI/ML models on the basis of at least one of a cell in which the communication apparatus is situated, a frequency band used by the communication apparatus, or an SSB used by the communication apparatus, and uses the selected one of the plurality of the AI/ML models. [18] The communication apparatus according to any one of [1] to [17], in which the signal processor includes a plurality of the AI/ML models, and the controller selects one of the plurality of the AI/ML models according to a radio-resource-control (RRC) state of the communication apparatus, and uses the selected one of the plurality of the AI/ML models. [19] The communication apparatus according to any one of [1] to [18], in which the communication apparatus uses the AI/ML model set on the basis of the information on which to base learning, the information being included in system information received from the other communication apparatus, or the communication apparatus uses the AI/ML model set and trained on the basis of the information on which to base learning, the information being included in the system information received from the other communication apparatus. when the communication apparatus is in an RRC idle state, [20] The communication apparatus according to [19], in which the communication apparatus uses the AI/ML model set on the basis of the information on which to base learning, the information being received from the other communication apparatus during an RRC connection, or the communication apparatus uses the AI/ML model set and trained on the basis of the information on which to base learning, the information being received from the other communication apparatus during the RRC connection, or the communication apparatus uses the AI/ML model set on the basis of the information on which to base learning, the information being included in system information received from the other communication apparatus, or the communication apparatus uses the AI/ML model set and trained on the basis of the information on which to base learning, the information being included in the system information received from the other communication apparatus. when the communication apparatus is in an RRC inactive state, [21] The communication apparatus according to [19] or [20], in which the other communication apparatus receives, from a data set providing apparatus, the information on which to base learning, the data set providing apparatus being accessible in common by communication apparatuses of different vendors. [22] The communication apparatus according to any one of [1] to [21], in which the receiver further receives, from the other communication apparatus, information necessary to access a data set providing apparatus that is accessible in common by communication apparatuses of different vendors. [23] The communication apparatus according to any one of [1] to [22], in which the other communication apparatus is a terminal apparatus, the communication apparatus is a base station that establishes wireless communication with the terminal apparatus, and the communication apparatus further includes a transmitter that transmits, to another base station, at least one of information regarding the AI/ML model having performed learning, the information on which to base learning of an AI/ML model, or capability information regarding the capability for an AI/ML model, the transmission being performed when the terminal apparatus is handed over from the base station to the other base station. [24] The communication apparatus according to any one of [1] to [23], in which a storage that stores therein information on which to base learning of an AI/ML model; a receiver that receives, from a communication apparatus, a request for provision of the information on which to base learning of an AI/ML model; and a transmitter that transmits, to the communication apparatus, the information on which to base learning of an AI/ML model. [25] A data set providing apparatus that is accessible in common by communication apparatuses of different vendors, the data set providing apparatus including: the data set providing apparatus is defined as a network function. [26] The data set providing apparatus according to [25], in which receiving information on which to base learning; and setting the AI/ML model on the basis of the information on which to base learning. [27] A method for training an AI/ML model, the method including: receiving a request for provision of the information on which to base learning; and transmitting the information on which to base learning to a transmission source that has transmitted the request for the provision. [28] A method for providing information on which to base learning of an AI/ML model, the method including: Further, the present disclosure may also take the following configurations.

1 wireless communication system 10 management apparatus 11 communication section 12 storage 13 controller 20 base station (communication apparatus, another communication apparatus) 21 wireless communication section 211 transmitter 212 receiver 213 antenna 22 storage 23 decoder (signal processor) 25 controller 30 relay station (communication apparatus, another communication apparatus) 31 wireless communication section 311 transmitter 312 receiver 313 antenna 32 storage 33 decoder (signal processor) 35 controller 40 terminal apparatus (communication apparatus, another communication apparatus) 41 wireless communication section 411 transmitter 412 receiver 413 antenna 42 storage 43 decoder (signal processor) 45 controller 50 data set providing apparatus 51 storage 52 receiver 53 transmitter